196 
PACIFIC SCIENCE, Vol. XVIII, April 1964 
comprises three small internal black spots that 
show through the somites. Roule and Bertin 
stated that on younger larvae each of the three 
spots is large, branched, and extends over the 
width of several somites. On the 147-mm larva 
in the Scripps collection, each spot consists of 
a loose cluster of from two to several cells. Roule 
and Bertin found that the spots become much 
less conspicuous on the larger larvae, but traces 
of them remain on some of the metamorphosing 
specimens. In gross appearance, the spots are 
farther forward on larger larvae than on small 
ones. This does not represent an actual displace- 
ment of the markings, however, but is a passive 
proportional modification that results from the 
continued addition and lengthening of somites 
posteriorly. The spots continue to occupy fixed 
positions (discussed in detail below). Roule and 
Bertin (op. cit., p. 74) found that in the ex- 
tensive "Dana” material the three spots occupy 
average locations at somites 39, 73, and 116 
respectively. 
Comparison with Leptocephalus acuticeps. 
The combination of excessively attenuated form 
and extremely high somite count tends to isolate 
larvae of Nemichthys from other leptocephali 
and to mask any similarity to them, but a critical 
examination of details reveals characters that 
link Nemichthys closely with Leptocephalus 
acuticeps . Direct comparison of these larvae 
shows that their most conspicuous differences 
correlate rather simply with their greatly dif- 
ferent somite counts and tail-tip structure. In 
Nemichthys the visceral anatomy borders a 
greater number of somites, so that the various 
organs are associated with more posterior so- 
mites than is true of their counterparts in L. 
acuticeps, but the visceral proportions are much 
the same as in acuticeps. The two kinds of 
larvae differ markedly in tail-tip structure. L . 
acuticeps has a bluntly rounded tail tip with a 
well-defined caudal fin; the Nemichthys larva 
has a thin filament-like tail tip with little or no 
apparent definition of fin elements. The Ne- 
michthys larva and L. acuticeps both have in- 
ternal rows of supraspinal and supraintestinal 
melanophores (which are relatively uncommon 
in eel larvae), and they agree in the small size 
and dense spacing of the cells in these rows. 
The most significant shared pigment character 
is the presence and similar placement of the 
unique pattern of three internal spots. It seems 
suprising that when Bertin (1936) redescribed 
the original larva of L. acuticeps he did not 
mention the resemblance of its internal three- 
spot pattern to that of Nemichthys , which he 
(with Roule) had described in detail only a 
few years before. Apparently, however, Bertin 
considered the extreme elongation and high 
somite counts of the Nemichthys larva to be 
primary characters of sufficient importance to 
outweigh any resemblances to other leptocephali. 
Perhaps, also, he had not yet studied a suf- 
ficiently wide variety of larval color patterns to 
realize the uniqueness of this one. L. acuticeps 
differs from Nemichthys in having melano- 
phores along the middorsal surface and in hav- 
ing more extensive pigment midventrally from 
the pericardium to the anus. Nemichthys appears 
to lack the scattered, isolated melanophores in 
the median connective-tissue zone which, in L. 
acuticeps, supplement the three aggregate spots. 
In respect to somite numbers, the three in- 
ternal spots are somewhat farther forward in L. 
acuticeps than in Nemichthys, but in relation to 
the total length the reverse condition eventually 
occurs because of the differences in the nature 
of the proportional changes during growth. Both 
kinds of larvae are subject to lengthening 
through enlargement of somites, but in addition 
Nemichthys lengthens through the continued 
formation of new somites posteriorly. Thus, in 
Nemichthys the posterior end literally grows 
away from the spot pattern. A superficial exami- 
nation of the spot positions reveals only that the 
two kinds of larvae differ, but a comparison 
of the spot positions with visceral "landmarks” 
instead of simply with somites reveals a striking 
agreement. Although the first spot averages 
about 10-14 somites farther back in Nemichthys 
than in the Scripps larvae of acuticeps, in both 
forms the spot occurs above or just behind the 
pylorus, close to or overlapping the position of 
a median vertical artery that extends down from 
the aorta to the viscera. The second spot av- 
erages about 20-23 somites farther back in 
Nemichthys than in acuticeps, but in both kinds 
the spot occupies the same morphological posi- 
tion about 10-12 somites ahead of the posterior 
end of the kidney, near or overlapping the 
